Weeks after ChatGPT-4 was released, an upgraded version, ChatGPT-5, is in the pipeline and expected to complete training by December, according to developer Siqi Chen.
But could the technology that supports ChatGPT have the ability to reason like a human being? It is an incredible claim. AI sentience is a hotly-debated topic that has yet to gain consensus within the research and developer community.
Sentient AI this year?
Writing on Twitter, Chen indicated the latest GPT is on its way. His remarks come only a few weeks after the release of ChatGPT-4. And it’s been less than six months since the initial release of ChatGPT in November 2022, which became an instant hit on the global market.
ChatGPT-4, which is still only available to ChatGPT Plus subscribers and not accessible to users in China and parts of Africa, has largely overshadowed its predecessor, the 3.5 model.
On his Tweeter feed, Chen said: “I have been told that GPT-5 is scheduled to complete training this December and that OpenAI expects it to achieve AGI, which means we will all hotly debate as to whether it actually achieves AGI.”
i have been told that gpt5 is scheduled to complete training this december and that openai expects it to achieve agi.
which means we will all hotly debate as to whether it actually achieves agi.
which means it will.
— Siqi Chen (@blader) March 27, 2023
AGI – Artificial General Intelligence – refers to the hypothetical ability of an AI system to perform any intellectual task that a human can do. This includes tasks such as reasoning, problem-solving, learning, perception, and natural language processing.
If ChatGPT-5 manages to reach AGI, then OpenAI would have hit a major milestone in generative AI. With the GPT-5 upgrade, it might be difficult to differentiate between a human’s response and that of the chatbot.
Given that most AI systems with AGI are made to perform one or a few tasks, ChatGPT-5 would become almost certainly the most developed AI tool.
OpenAI certainly contend that ChatGPT-5 will be the most advanced version of the GPT language developed to date. It will understand and generate natural language in a more sophisticated and nuanced way than ever before, they say.
The model will push the boundaries of what’s possible with machine learning, and it has the potential to revolutionize the way humans communicate and interact with technology.
A blessing and a curse
With human-like reasoning capabilities, ChatGPT might become a powerful tool capable of crafting nearly perfect content. But it could also be dangerous.
According to writer Alex Blake, bestowing an AI with that much power could have unintended consequences, ones that the world simply hasn’t thought of yet.
“It doesn’t mean the robot apocalypse is imminent, but it certainly raises a lot of questions about what the negative effects of AGI could be,” writes Blake.
Blake further explains that if the AGI goes off rails, it could enable the spread of incredibly convincing bots on social media channels like Twitter, helping to disseminate harmful disinformation and propaganda that is difficult to spot.
Twitter boss Elon Musk has supposedly been aware of this and made fighting AI bots “a key pillar of his tenure as the Twitter CEO.” Indeed, Musk recently co-signed an open letter calling for a pause to AI development.
GPT-4.5 to lay the ground for ChatGPT-5
Despite the Twitter comments, there is no official word from OpenAI with regards to ChatGPT-5, except to say that ChatGPT-4.5 will be “introduced in September or October 2023 as an intermediate version between the GPT-4 and the upcoming GPT-5.”
According to the company, future advancements in GPT-4.5 will likely serve as a solid foundation for the innovations expected from ChatGPT-5.
“By addressing the limitations of GPT-4 and introducing new improvements GPT-4.5 will play a crucial role in shaping the development of ChatGPT-5,” said OpenAI in a statement.
But users have not stopped speculating about the new tool. One, Moe Abbas, noted how ChatGPT is far away from reaching AGI.
— moeabbas (@realmoeabbas) March 27, 2023
“Altman said to Lex last week he doesn’t think that will happen until gpt-10,” another user, Judicialreview replied on the same thread.
Whether the coming upgrade achieves AGI or not, it will likely be superior at content generation, reasoning, and suffer from less “hallucination” than its predecessor, ChatGPT-4.
The AI chatbot has already racked up millions of users, and caused controversy in arts and education with some schools and universities banning its use.